Showing posts with label tags. Show all posts
Showing posts with label tags. Show all posts

Monday, August 21, 2006

Face and Image recognition for the masses

Ever wanted to search your entire image archive by typing `Peter' in the search box, and all the pictures with Peter's face in them turn up? Then, enter`Steffi Graf' in the text search box and all her pictures on the web turn up? Or, just upload a picture of a piece of jewelry to a site and it'll search the web for prices of this piece?

Riya.com comes a little closer to this dream. It's an introduction to what google might do with their new acquisition - Neven Vision - a company good at extracting information from photos. Google WAS talking to riya.com before, but I guess it didn't work out.

Discover Magazine has a good writeup on Riya here. Yeah, I still read some nerd print stuff. Coz the authors are more nerdy than some of the web-only types.

Anyway, riya.com just asks you to upload some pictures of your own to their website via a 43MB picture uploader software, and they'll recognize faces, objects, etc. in your photos. Then you can tag them in a number of ways, or riya will autotag them using information from other riya users who appeared in your photos.

Normally, I'd really scream about a site with features as powerful and unique like Riya's. But photos - wow! They're really personal, and people could do bad things with them. To top it all, I've a bad feeling about Riya - check out the bitchiness of Riya's CEO in his personal blog after Google bought Neven Vision instead of Riya. And note the `sour grapes' and incredibly bitchy comments made by the same CEO about Neven Vision and Google. With such bitchiness, I don't know whether I can trust the privacy statements or not. I don't even know whether this company has a future or not. A future in India, maybe, but a global future needs a broader perspective, I think, not the bitchiness demonstrated.

And another stupid, scary thing is, to use one of the modes of autotagging, it requires you to enter your google/yahoo/etc password and username to extract user information and contacts. Come on! That is disrespect of the crudest sense. They could just have an IMPORT CONTACTS function, instead of asking you for one of the most sacred things, your password to your gmail account and of course, your entire google accounts universe, including your credit card number in your checkout account and your personal search information.

So, after they asked for my google username and password, I decided to stop my picture uploading to Riya before it began. I've gotta wait till google's debut of this type of technology. And between google and a bitchy CEO, I'd trust google anytime if I had to.

The bottom line - How much does image recognition sharpen your axe? A whole lot, I think - tagging my 12,000 or more photos, is an IMPOSSIBLE task. With computer-aided tagging, I'd do it. And once the tagging is done, based on facial and object recognition, most of the new pictures would be tagged automatically based on previous object recognition data, unless I've got a new baby or new friends in the pictures. With google's network, perhaps some of your friends have elected to share their own face recognition metrics with every one of their gmail contacts, and I may not even have to manually tag many of my new friends.

If you're a teacher and you'd have to take attendance, just snap away with your point and shoot, 3-4 pictures, and upload it, and there, all 30 of your students identified (unless someone made a funny face, but if he makes funny faces all the time, he'd be recognized too).

And if you found a nice boy and wanted to see how nice this boy really was, use your new K800, take a picture, and see whether this boy appears in some slut's blog as a `good partner'. Or, whether he appears in a picture of a Taliban training camp.

Obviously, with this, nobody can hide, nobody can run. It's pretty much beyond your control - photos usually are taken by friends, and you'd probably have appeared in photos before. Chances are a few of your friends upload the picture, and there - Mr. X identified - in name, with photos.

Doesn't bother me that much. Just hope that Google's CEO isn't as bitchy as Riya's.

Wednesday, August 16, 2006

Abraham Lincoln and his axe

A vastly underrated quote of Abraham Lincoln's: "If I had six hours to chop down a tree, I'd spend the first four hours sharpening the axe".

Just to state the obvious, the axe is a tool to perform what you have to do. Since the entire human race, including those living in underdeveloped states like Myanmar, has shifted in varying degrees to the information age, I can probably define the axe as `something which helps me do what I have to do faster, better and cheaper'.

I spend an inordinate amount of time sharpening my axe. In my first year at law school, I sharpened my axe till the eve of my exams for Criminal Law, and flunked it. At work, I spend almost all my time sharpening my axe.

But now, with websites like Lifehacker and a ton of like `productivity blogs' around, I find like minded souls, and I spend even more time, if it's possible to sharpen my axe. Trouble is, there seems to be almost no limit to how sharp the axe can be.

So now, I can do a whole lot of things much faster, better and easier than many of my contemporaries. I google faster, I use the spreadsheet faster, I populate the spreadsheet from the database faster. Drawing resources from experts like programmers, and other hackers, and a veritable network of friends and `productivity bloggers', the rate of sharpening is speeding up.

So much so that, in the last 6 months, I feel the gulf between the guys with sharp axes and less keen blades, has widened tremendously. More than it has for the last 3 years. I have no concrete reasons on `why' but I speculate it's because Eric Schmidt, the boss of Google, works harder than Larry Page and Sergey Brin.

Yes, in my mind, Google has driven the industry and provided it with all the catalysts for growth. Google has, by design, taken the stuffy corporate spin out of the web, and replaced it with fun. Witness their blogs - it goes something like:

Well, Google XXX has always prided itself on 2 principles:

1) Our integration of XXX with our great search engine, synergies (blah blah serious corporate stuff) and
2) Our love for jellybeans in Oreo cookies mashed with Ice Cream (fun stuff)

It's their great expertise at the mind game which is screwing Microsoft more than any other. And it's screwing all those suits at Oracle too. Make that, suits everywhere. What the hell do you wear a suit for if a bunch of fun loving kids at Mountain View earn US$800M a quarter and have fun at it? They put the emphasis on STUFFY in stuffy suits.

I'm not here to judge how genuine the fun is at google. What I have to say is that they're doing a damn good job. They put the T-shirts back at Silicon Valley.

Oh and why am I talking about google? Because, the web services which google made so powerful and attractive and fun, has done a helluva lot to sharpen my axe lately. Searches have become better. People are dying to beat google at the game, and labels and tags have given information a whole new dimension. Countless new dimensions. Information quality has improved greatly just because of the new ways to categorize it.

The bottom line: I got this blog to test Blogger Beta, but they didn't allow me to transfer my other blogs to the beta yet, and forced me to register a new name. Since I had this in mind for a long time already, almost 3 months, so I decided to register a new blog.

I don't know whether I have time to fill in this blog past this introduction, but if I do, I hope you get something outta it. I'm gonna show you how I sharpened my axe.